Abstract
We consider the situation when treatments are the combinations of levels of two factors, one of which exerts competition effects. Experimental units consist of one or several rows of plots with border plots. Classes of balanced designs which are optimal for the estimation of direct treatment effects and competition effects are introduced, and methods of construction are given for various numbers of levels of each factor.
| Original language | English |
|---|---|
| Pages (from-to) | 643-659 |
| Number of pages | 17 |
| Journal | Biometrika |
| Volume | 80 |
| Issue number | 3 |
| Publication status | Published - 1993 |
Keywords
- competition
- neighbour-balanced factorial design
- optimal design